I believe this model to be very good at photorealism although it does have some artistic capabilities as well. Here are a few tricks to using the SDXL versions of GonzaLomo to its fullest potential:
Clip Skip is your friend. I recommend trying a prompt with both 1 and 2 values. 1 can make images that are more creative, artistic, retro or grungy whereas 2 is better at photorealism and the, um, physical attractiveness of the subject(s).
Generate at 1024x1536 or higher using the PatchModelAddDownscale node in ComfyUI. This will create a much more detailed image that can then be upscaled further, resulting in incredible photorealistic detail. Using that node, I can do an initial generation as high as 1280x1920.
You can play around with different sampler settings but I find that the absolute best results are with LCM/Karras, 12 steps and CFG 1.0.
Selective usage of the Self Attention Guidance node can sometimes give amazing results. I always have it in the pipeline and usually bypassed but turn it on when I think an image has potential but just needs a little more oomph like color saturation or contrast.
I get really interesting/creative results by generating with the V1.1 Semi-Real model and then resampling that image with the V2.0 Unity model at around 0.60 denoise. You really need to play with the denoise to find the right value that tips the image from cartoonish to photorealism but you'll be amazed at what's possible when you do. Note that V1.1 Semi-Real requires Clip Skip 2+ and will not work with 1 at all.
And lastly, always be upscaling! More pixels equals more detail equals more photorealism.
Note: I only use ComfyUI so I apologize in advance if you're having trouble with my model in some other tool and I can't help.
Current Models
V2.0 Unity
Essentially, this is V1.1 Real with a bit of Siren Forge, Cyberrealistic and Lustify added to the mix. The base is still Gonzales and LomoXL.
I called this version Unity because I originally wanted to have separate models for photorealism and artistic styles. However, I realized that I could get both in one model and created V1.1 Artsy for that purpose but it was a failure. This is my second attempt at that combination and I think/hope it's a success.
At the very least, I don't want to degrade the photorealism capabilities and just want to add more surreal, artistic or fantastical styles as an option.
The recommended settings haven't changed since V1.1 Real:
Sampler: LCM (also dpmpp2a and ddim can provide interesting results)
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 1-5
I normally keep my settings at LCM/Karras with a CFG of 1.0 but I tend to vary the clip skip between 1 and 5 because those can change the image pretty drastically. Sometimes an image at clip skip 2 will be terrible but might become amazing with clip skip 1, and vice-versa.
V1.1 Semi-Real
I'm doing a little experimenting with non-photorealistic models. This is a merge of GonzaLomo v1.1 Real and ToonComix which is a toon model that does a really good job with detailed backgrounds. The whole Comix family of models is very good so check them out.
Sampler: LCM
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 2-5 (1 is not supported)
Retired Models
V1.1 Real
This is a simple 60/40 merge between two of my favorite photorealistic models - Gonzales-NSFW-PonyV1/V2-DMD v2.0 and LomoXL (with a little bit of Lustify thrown in). Many, many thanks to the makers of those excellent models.
Gonzales is a Pony model so GonzaLomo inherits all of those capabilties. LomoXL has a very strong analog feel to it which helps give an edge to Gonzales' "pony-ness".
Sampler: LCM
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 1 allows greater analog effect but lessens NSFW capabilities
Clip Skip: 2-5 lessens/removes the analog effect but allows more NSFW
Try this workflow to get started, if you need it.
To be clear, I created this merge for my own use cases. It's certainly not better than Gonzales for most things. I'm just hoping to inject Gonzales with some of the gritty, analog feel that I was looking for in some cases. LomoXL also allows for a wider variation of faces, especially with clip skip 1.
Enjoy!
P.S. This is my first model upload so I would appreciate any constructive feedback or comments.
V1.1 Artsy alpha (Not very good imo, use V2.0 Unity instead)
This was an attempted merge between V1.1 Real and a few art-focused models. The main purpose of this model is to be able to produce artistic styles but it has the tendency to try to force photorealism when prompted clearly for an artistic style. This model has been abandoned in favor of V2.0 Unity which does a much better job of combining the photorealism of V1.1 Real with artistic styles.
Description
Thought I'd try my hand at a 100% Flux Schnell version. A light refinement with an SDXL model is recommended but not necessary. The images in the showcase post for this model were generated using only this model. All other images will be generated using my Flux Refiner Workflow which refines the images using my SDXL model.
Steps: 3-6
Flux Guidance: Doesn't matter because Schnell ignores this.
Sampler: dpmpp_2m, dpm_2, euler
Scheduler: simple, beta
CFG: 1.0-1.3
This model can be used like a regular checkpoint as it contains both the Flux VAE and CLIP (t5xxl_fp8_e4m3fn and ViT-L-14-TEXT-detail-improved-hiT-GmP-TE-only-HF).
FAQ
Comments (14)
Once again you have changed my working ways. I use Fooocus but it doesn't support Flux. Today I tried your new Flux checkpoint with your refiner workflow in Comfyui and the results are hugely impressive. Thank you again.
Awesome! I'm glad the model and workflow are working so well for you. As you're someone moving from Fooocus to ComfyUI, I'd appreciate any other feedback you have, if any. I'd like to make it as beginner friendly as possible.
@GBRX I'm not used to creating my own workflows for Comfyui so I rely on downloads such as yours (thanks again). I became frustrated with Fooocus because I wanted Flux quality but was restricted to SDXL and Pony. However your recent models improved the quality of each of those so much so that I forgot about Flux. Now though, you have come through again with a Flux model, albeit in Comfy. I'm fine with that although I'm using your settings because I wouldn't know what to change. All I did was add Power Lora Loader (rgthree) and that works fine so far. I might play with different samplers but I like the results I am getting right now.
@dcham2310 The settings you're going to want to play around with most are the Flux steps, sampler and resolution. Adjusting those will have the biggest impact on the base image.
The steps should be between 3 and 6. Lower than 3 and you will lose too much detail. Higher than 6 and you'll either see no benefit or too much of a good thing.
I think dpmpp_2m works ideally with this model, providing the best color and composition. However, dpm_2 is also good and euler is fine. The scheduler is best left at simple but beta sometimes works a little better.
The resolutionn can be whatever you want but changing it and keeping the same seed/prompt can either greatly improve an image or ruin it.
On the SDXL refining side, the only setting you probably want to change is the denoise. You want the denoise to be just high enough to layer on some detail without making drastic changes to the base image. I find between 0.25 and 0.4 is best. The goal is to marry the excellent color and composition that Flux provides with the superior textures of SDXL.
@GBRX Thanks for the tips. One last request: do you have any Comfyui workflows for your other models? Now I've had a taste I want to compare results with the Fooocus/SDXL/Pony models. As I said, I'm not experienced enough with Comfy to know what nodes to use with which model, etc.
@dcham2310 Yes, I have this workflow for the SDXL models, https://civitai.com/models/1513601?modelVersionId=1875485. It works a lot like the Flux workflow except that the base generation uses the SDXL model and a couple different nodes. However, I just took a look at both my Flux and SDXL workflows and it looks like they need some updating. They're fine for you to use as-is but I'm going to be putting out new versions tonight or tomorrow.
@dcham2310 I've uploaded a v3 version of the Flux Refiner Workflow if you want to take a look. https://civitai.com/models/1651835?modelVersionId=1974042
@GBRX Thanks, the notes are very helpful. Playing with different checkpoints (SDXL section) can make significant changes. I tried a specialist refiner and the results were awful (edit: I needed to change the sampler and then the refiner worked ok). Yours works well as does a CyberRealistic model. The denoise in that section is very sensitive and small changes can be the difference between excellent and unacceptable.
@dcham2310 Yeah, the denoise is probably the one setting that I have to mess with the most. As far as other SDXL checkpoints, there's a ton of great ones these days - XXX Ray, Analog Madness, Epic Realism, Big Love, Gonzales, CyberRealistic, Lustify. More than I can remember right now. Try them all.
Some of the Flux images you posted look awesome, but skin still looks like plastic too often and there are some slightly visible vertical and horizontal banding on many of the pictures. That's not your fault though. It's been a problem with Flux for a long time.
The showcase images for the model were created using the Flux model alone, so those are going to be the lowest quality. Many of them definitely still have Flux skin and the banding that you mentioned. However, I think it's important for a model to showcase what it's solely capable of before any other refinement is done. That said, all the rest of the images I created with the model use my refiner workflow which refines the Flux image with my SDXL model and layers on skin details as it upscales the image. If any of those still look a little Fluxy then it's because I didn't use a high enough denoise during the refinement. I can be lazy like that.
very good
Flux is interesting idea. I think it would be good as Chroma LORA, or Chroma merge (probably possible).
Thanks, I still need to investigate Chroma.